Reaction force inspection system using regression analysis and neural network

This paper looks at a system for inspection of the quality of a probe's reaction force characteristics. This system automates the inspection method utilizing the touching of an inspector's finger. Regression analysis and a neural network approach are applied to the system for probes to learn an inspector's Finger judgment. From experimental results, the effectiveness of the system has been clarified. in this paper, an inspection method is described for the analysis of the quality of a probe's reaction force characteristics. Its aim is to classify reaction force signals, as the recognition of such signals is ofparticular interest in the field of inspection5'. The classification we perform is based on regression analysis and a neural network, an approach needed for inspections by probes in real time. The inspection system evaluates the operative feeling of probes. Experimental results show that this method is suitable for such inspections.